xref: /freebsd/sys/rpc/svc_vc.c (revision 47dd1d1b619cc035b82b49a91a25544309ff95ae)
1 /*	$NetBSD: svc_vc.c,v 1.7 2000/08/03 00:01:53 fvdl Exp $	*/
2 
3 /*-
4  * SPDX-License-Identifier: BSD-3-Clause
5  *
6  * Copyright (c) 2009, Sun Microsystems, Inc.
7  * All rights reserved.
8  *
9  * Redistribution and use in source and binary forms, with or without
10  * modification, are permitted provided that the following conditions are met:
11  * - Redistributions of source code must retain the above copyright notice,
12  *   this list of conditions and the following disclaimer.
13  * - Redistributions in binary form must reproduce the above copyright notice,
14  *   this list of conditions and the following disclaimer in the documentation
15  *   and/or other materials provided with the distribution.
16  * - Neither the name of Sun Microsystems, Inc. nor the names of its
17  *   contributors may be used to endorse or promote products derived
18  *   from this software without specific prior written permission.
19  *
20  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
21  * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23  * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
24  * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
25  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
26  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
27  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
28  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
29  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
30  * POSSIBILITY OF SUCH DAMAGE.
31  */
32 
33 #if defined(LIBC_SCCS) && !defined(lint)
34 static char *sccsid2 = "@(#)svc_tcp.c 1.21 87/08/11 Copyr 1984 Sun Micro";
35 static char *sccsid = "@(#)svc_tcp.c	2.2 88/08/01 4.0 RPCSRC";
36 #endif
37 #include <sys/cdefs.h>
38 __FBSDID("$FreeBSD$");
39 
40 /*
41  * svc_vc.c, Server side for Connection Oriented based RPC.
42  *
43  * Actually implements two flavors of transporter -
44  * a tcp rendezvouser (a listner and connection establisher)
45  * and a record/tcp stream.
46  */
47 
48 #include <sys/param.h>
49 #include <sys/lock.h>
50 #include <sys/kernel.h>
51 #include <sys/malloc.h>
52 #include <sys/mbuf.h>
53 #include <sys/mutex.h>
54 #include <sys/proc.h>
55 #include <sys/protosw.h>
56 #include <sys/queue.h>
57 #include <sys/socket.h>
58 #include <sys/socketvar.h>
59 #include <sys/sx.h>
60 #include <sys/systm.h>
61 #include <sys/uio.h>
62 
63 #include <net/vnet.h>
64 
65 #include <netinet/tcp.h>
66 
67 #include <rpc/rpc.h>
68 
69 #include <rpc/krpc.h>
70 #include <rpc/rpc_com.h>
71 
72 #include <security/mac/mac_framework.h>
73 
74 static bool_t svc_vc_rendezvous_recv(SVCXPRT *, struct rpc_msg *,
75     struct sockaddr **, struct mbuf **);
76 static enum xprt_stat svc_vc_rendezvous_stat(SVCXPRT *);
77 static void svc_vc_rendezvous_destroy(SVCXPRT *);
78 static bool_t svc_vc_null(void);
79 static void svc_vc_destroy(SVCXPRT *);
80 static enum xprt_stat svc_vc_stat(SVCXPRT *);
81 static bool_t svc_vc_ack(SVCXPRT *, uint32_t *);
82 static bool_t svc_vc_recv(SVCXPRT *, struct rpc_msg *,
83     struct sockaddr **, struct mbuf **);
84 static bool_t svc_vc_reply(SVCXPRT *, struct rpc_msg *,
85     struct sockaddr *, struct mbuf *, uint32_t *seq);
86 static bool_t svc_vc_control(SVCXPRT *xprt, const u_int rq, void *in);
87 static bool_t svc_vc_rendezvous_control (SVCXPRT *xprt, const u_int rq,
88     void *in);
89 static void svc_vc_backchannel_destroy(SVCXPRT *);
90 static enum xprt_stat svc_vc_backchannel_stat(SVCXPRT *);
91 static bool_t svc_vc_backchannel_recv(SVCXPRT *, struct rpc_msg *,
92     struct sockaddr **, struct mbuf **);
93 static bool_t svc_vc_backchannel_reply(SVCXPRT *, struct rpc_msg *,
94     struct sockaddr *, struct mbuf *, uint32_t *);
95 static bool_t svc_vc_backchannel_control(SVCXPRT *xprt, const u_int rq,
96     void *in);
97 static SVCXPRT *svc_vc_create_conn(SVCPOOL *pool, struct socket *so,
98     struct sockaddr *raddr);
99 static int svc_vc_accept(struct socket *head, struct socket **sop);
100 static int svc_vc_soupcall(struct socket *so, void *arg, int waitflag);
101 static int svc_vc_rendezvous_soupcall(struct socket *, void *, int);
102 
103 static struct xp_ops svc_vc_rendezvous_ops = {
104 	.xp_recv =	svc_vc_rendezvous_recv,
105 	.xp_stat =	svc_vc_rendezvous_stat,
106 	.xp_reply =	(bool_t (*)(SVCXPRT *, struct rpc_msg *,
107 		struct sockaddr *, struct mbuf *, uint32_t *))svc_vc_null,
108 	.xp_destroy =	svc_vc_rendezvous_destroy,
109 	.xp_control =	svc_vc_rendezvous_control
110 };
111 
112 static struct xp_ops svc_vc_ops = {
113 	.xp_recv =	svc_vc_recv,
114 	.xp_stat =	svc_vc_stat,
115 	.xp_ack =	svc_vc_ack,
116 	.xp_reply =	svc_vc_reply,
117 	.xp_destroy =	svc_vc_destroy,
118 	.xp_control =	svc_vc_control
119 };
120 
121 static struct xp_ops svc_vc_backchannel_ops = {
122 	.xp_recv =	svc_vc_backchannel_recv,
123 	.xp_stat =	svc_vc_backchannel_stat,
124 	.xp_reply =	svc_vc_backchannel_reply,
125 	.xp_destroy =	svc_vc_backchannel_destroy,
126 	.xp_control =	svc_vc_backchannel_control
127 };
128 
129 /*
130  * Usage:
131  *	xprt = svc_vc_create(sock, send_buf_size, recv_buf_size);
132  *
133  * Creates, registers, and returns a (rpc) tcp based transporter.
134  * Once *xprt is initialized, it is registered as a transporter
135  * see (svc.h, xprt_register).  This routine returns
136  * a NULL if a problem occurred.
137  *
138  * The filedescriptor passed in is expected to refer to a bound, but
139  * not yet connected socket.
140  *
141  * Since streams do buffered io similar to stdio, the caller can specify
142  * how big the send and receive buffers are via the second and third parms;
143  * 0 => use the system default.
144  */
145 SVCXPRT *
146 svc_vc_create(SVCPOOL *pool, struct socket *so, size_t sendsize,
147     size_t recvsize)
148 {
149 	SVCXPRT *xprt;
150 	struct sockaddr* sa;
151 	int error;
152 
153 	SOCK_LOCK(so);
154 	if (so->so_state & (SS_ISCONNECTED|SS_ISDISCONNECTED)) {
155 		SOCK_UNLOCK(so);
156 		CURVNET_SET(so->so_vnet);
157 		error = so->so_proto->pr_usrreqs->pru_peeraddr(so, &sa);
158 		CURVNET_RESTORE();
159 		if (error)
160 			return (NULL);
161 		xprt = svc_vc_create_conn(pool, so, sa);
162 		free(sa, M_SONAME);
163 		return (xprt);
164 	}
165 	SOCK_UNLOCK(so);
166 
167 	xprt = svc_xprt_alloc();
168 	sx_init(&xprt->xp_lock, "xprt->xp_lock");
169 	xprt->xp_pool = pool;
170 	xprt->xp_socket = so;
171 	xprt->xp_p1 = NULL;
172 	xprt->xp_p2 = NULL;
173 	xprt->xp_ops = &svc_vc_rendezvous_ops;
174 
175 	CURVNET_SET(so->so_vnet);
176 	error = so->so_proto->pr_usrreqs->pru_sockaddr(so, &sa);
177 	CURVNET_RESTORE();
178 	if (error) {
179 		goto cleanup_svc_vc_create;
180 	}
181 
182 	memcpy(&xprt->xp_ltaddr, sa, sa->sa_len);
183 	free(sa, M_SONAME);
184 
185 	xprt_register(xprt);
186 
187 	solisten(so, -1, curthread);
188 
189 	SOLISTEN_LOCK(so);
190 	xprt->xp_upcallset = 1;
191 	solisten_upcall_set(so, svc_vc_rendezvous_soupcall, xprt);
192 	SOLISTEN_UNLOCK(so);
193 
194 	return (xprt);
195 
196 cleanup_svc_vc_create:
197 	sx_destroy(&xprt->xp_lock);
198 	svc_xprt_free(xprt);
199 
200 	return (NULL);
201 }
202 
203 /*
204  * Create a new transport for a socket optained via soaccept().
205  */
206 SVCXPRT *
207 svc_vc_create_conn(SVCPOOL *pool, struct socket *so, struct sockaddr *raddr)
208 {
209 	SVCXPRT *xprt;
210 	struct cf_conn *cd;
211 	struct sockaddr* sa = NULL;
212 	struct sockopt opt;
213 	int one = 1;
214 	int error;
215 
216 	bzero(&opt, sizeof(struct sockopt));
217 	opt.sopt_dir = SOPT_SET;
218 	opt.sopt_level = SOL_SOCKET;
219 	opt.sopt_name = SO_KEEPALIVE;
220 	opt.sopt_val = &one;
221 	opt.sopt_valsize = sizeof(one);
222 	error = sosetopt(so, &opt);
223 	if (error) {
224 		return (NULL);
225 	}
226 
227 	if (so->so_proto->pr_protocol == IPPROTO_TCP) {
228 		bzero(&opt, sizeof(struct sockopt));
229 		opt.sopt_dir = SOPT_SET;
230 		opt.sopt_level = IPPROTO_TCP;
231 		opt.sopt_name = TCP_NODELAY;
232 		opt.sopt_val = &one;
233 		opt.sopt_valsize = sizeof(one);
234 		error = sosetopt(so, &opt);
235 		if (error) {
236 			return (NULL);
237 		}
238 	}
239 
240 	cd = mem_alloc(sizeof(*cd));
241 	cd->strm_stat = XPRT_IDLE;
242 
243 	xprt = svc_xprt_alloc();
244 	sx_init(&xprt->xp_lock, "xprt->xp_lock");
245 	xprt->xp_pool = pool;
246 	xprt->xp_socket = so;
247 	xprt->xp_p1 = cd;
248 	xprt->xp_p2 = NULL;
249 	xprt->xp_ops = &svc_vc_ops;
250 
251 	/*
252 	 * See http://www.connectathon.org/talks96/nfstcp.pdf - client
253 	 * has a 5 minute timer, server has a 6 minute timer.
254 	 */
255 	xprt->xp_idletimeout = 6 * 60;
256 
257 	memcpy(&xprt->xp_rtaddr, raddr, raddr->sa_len);
258 
259 	CURVNET_SET(so->so_vnet);
260 	error = so->so_proto->pr_usrreqs->pru_sockaddr(so, &sa);
261 	CURVNET_RESTORE();
262 	if (error)
263 		goto cleanup_svc_vc_create;
264 
265 	memcpy(&xprt->xp_ltaddr, sa, sa->sa_len);
266 	free(sa, M_SONAME);
267 
268 	xprt_register(xprt);
269 
270 	SOCKBUF_LOCK(&so->so_rcv);
271 	xprt->xp_upcallset = 1;
272 	soupcall_set(so, SO_RCV, svc_vc_soupcall, xprt);
273 	SOCKBUF_UNLOCK(&so->so_rcv);
274 
275 	/*
276 	 * Throw the transport into the active list in case it already
277 	 * has some data buffered.
278 	 */
279 	sx_xlock(&xprt->xp_lock);
280 	xprt_active(xprt);
281 	sx_xunlock(&xprt->xp_lock);
282 
283 	return (xprt);
284 cleanup_svc_vc_create:
285 	sx_destroy(&xprt->xp_lock);
286 	svc_xprt_free(xprt);
287 	mem_free(cd, sizeof(*cd));
288 
289 	return (NULL);
290 }
291 
292 /*
293  * Create a new transport for a backchannel on a clnt_vc socket.
294  */
295 SVCXPRT *
296 svc_vc_create_backchannel(SVCPOOL *pool)
297 {
298 	SVCXPRT *xprt = NULL;
299 	struct cf_conn *cd = NULL;
300 
301 	cd = mem_alloc(sizeof(*cd));
302 	cd->strm_stat = XPRT_IDLE;
303 
304 	xprt = svc_xprt_alloc();
305 	sx_init(&xprt->xp_lock, "xprt->xp_lock");
306 	xprt->xp_pool = pool;
307 	xprt->xp_socket = NULL;
308 	xprt->xp_p1 = cd;
309 	xprt->xp_p2 = NULL;
310 	xprt->xp_ops = &svc_vc_backchannel_ops;
311 	return (xprt);
312 }
313 
314 /*
315  * This does all of the accept except the final call to soaccept. The
316  * caller will call soaccept after dropping its locks (soaccept may
317  * call malloc).
318  */
319 int
320 svc_vc_accept(struct socket *head, struct socket **sop)
321 {
322 	struct socket *so;
323 	int error = 0;
324 	short nbio;
325 
326 	/* XXXGL: shouldn't that be an assertion? */
327 	if ((head->so_options & SO_ACCEPTCONN) == 0) {
328 		error = EINVAL;
329 		goto done;
330 	}
331 #ifdef MAC
332 	error = mac_socket_check_accept(curthread->td_ucred, head);
333 	if (error != 0)
334 		goto done;
335 #endif
336 	/*
337 	 * XXXGL: we want non-blocking semantics.  The socket could be a
338 	 * socket created by kernel as well as socket shared with userland,
339 	 * so we can't be sure about presense of SS_NBIO.  We also shall not
340 	 * toggle it on the socket, since that may surprise userland.  So we
341 	 * set SS_NBIO only temporarily.
342 	 */
343 	SOLISTEN_LOCK(head);
344 	nbio = head->so_state & SS_NBIO;
345 	head->so_state |= SS_NBIO;
346 	error = solisten_dequeue(head, &so, 0);
347 	head->so_state &= (nbio & ~SS_NBIO);
348 	if (error)
349 		goto done;
350 
351 	so->so_state |= nbio;
352 	*sop = so;
353 
354 	/* connection has been removed from the listen queue */
355 	KNOTE_UNLOCKED(&head->so_rdsel.si_note, 0);
356 done:
357 	return (error);
358 }
359 
360 /*ARGSUSED*/
361 static bool_t
362 svc_vc_rendezvous_recv(SVCXPRT *xprt, struct rpc_msg *msg,
363     struct sockaddr **addrp, struct mbuf **mp)
364 {
365 	struct socket *so = NULL;
366 	struct sockaddr *sa = NULL;
367 	int error;
368 	SVCXPRT *new_xprt;
369 
370 	/*
371 	 * The socket upcall calls xprt_active() which will eventually
372 	 * cause the server to call us here. We attempt to accept a
373 	 * connection from the socket and turn it into a new
374 	 * transport. If the accept fails, we have drained all pending
375 	 * connections so we call xprt_inactive().
376 	 */
377 	sx_xlock(&xprt->xp_lock);
378 
379 	error = svc_vc_accept(xprt->xp_socket, &so);
380 
381 	if (error == EWOULDBLOCK) {
382 		/*
383 		 * We must re-test for new connections after taking
384 		 * the lock to protect us in the case where a new
385 		 * connection arrives after our call to accept fails
386 		 * with EWOULDBLOCK.
387 		 */
388 		SOLISTEN_LOCK(xprt->xp_socket);
389 		if (TAILQ_EMPTY(&xprt->xp_socket->sol_comp))
390 			xprt_inactive_self(xprt);
391 		SOLISTEN_UNLOCK(xprt->xp_socket);
392 		sx_xunlock(&xprt->xp_lock);
393 		return (FALSE);
394 	}
395 
396 	if (error) {
397 		SOLISTEN_LOCK(xprt->xp_socket);
398 		if (xprt->xp_upcallset) {
399 			xprt->xp_upcallset = 0;
400 			soupcall_clear(xprt->xp_socket, SO_RCV);
401 		}
402 		SOLISTEN_UNLOCK(xprt->xp_socket);
403 		xprt_inactive_self(xprt);
404 		sx_xunlock(&xprt->xp_lock);
405 		return (FALSE);
406 	}
407 
408 	sx_xunlock(&xprt->xp_lock);
409 
410 	sa = NULL;
411 	error = soaccept(so, &sa);
412 
413 	if (error) {
414 		/*
415 		 * XXX not sure if I need to call sofree or soclose here.
416 		 */
417 		if (sa)
418 			free(sa, M_SONAME);
419 		return (FALSE);
420 	}
421 
422 	/*
423 	 * svc_vc_create_conn will call xprt_register - we don't need
424 	 * to do anything with the new connection except derefence it.
425 	 */
426 	new_xprt = svc_vc_create_conn(xprt->xp_pool, so, sa);
427 	if (!new_xprt) {
428 		soclose(so);
429 	} else {
430 		SVC_RELEASE(new_xprt);
431 	}
432 
433 	free(sa, M_SONAME);
434 
435 	return (FALSE); /* there is never an rpc msg to be processed */
436 }
437 
438 /*ARGSUSED*/
439 static enum xprt_stat
440 svc_vc_rendezvous_stat(SVCXPRT *xprt)
441 {
442 
443 	return (XPRT_IDLE);
444 }
445 
446 static void
447 svc_vc_destroy_common(SVCXPRT *xprt)
448 {
449 
450 	if (xprt->xp_socket)
451 		(void)soclose(xprt->xp_socket);
452 
453 	if (xprt->xp_netid)
454 		(void) mem_free(xprt->xp_netid, strlen(xprt->xp_netid) + 1);
455 	svc_xprt_free(xprt);
456 }
457 
458 static void
459 svc_vc_rendezvous_destroy(SVCXPRT *xprt)
460 {
461 
462 	SOLISTEN_LOCK(xprt->xp_socket);
463 	if (xprt->xp_upcallset) {
464 		xprt->xp_upcallset = 0;
465 		solisten_upcall_set(xprt->xp_socket, NULL, NULL);
466 	}
467 	SOLISTEN_UNLOCK(xprt->xp_socket);
468 
469 	svc_vc_destroy_common(xprt);
470 }
471 
472 static void
473 svc_vc_destroy(SVCXPRT *xprt)
474 {
475 	struct cf_conn *cd = (struct cf_conn *)xprt->xp_p1;
476 
477 	SOCKBUF_LOCK(&xprt->xp_socket->so_rcv);
478 	if (xprt->xp_upcallset) {
479 		xprt->xp_upcallset = 0;
480 		soupcall_clear(xprt->xp_socket, SO_RCV);
481 	}
482 	SOCKBUF_UNLOCK(&xprt->xp_socket->so_rcv);
483 
484 	svc_vc_destroy_common(xprt);
485 
486 	if (cd->mreq)
487 		m_freem(cd->mreq);
488 	if (cd->mpending)
489 		m_freem(cd->mpending);
490 	mem_free(cd, sizeof(*cd));
491 }
492 
493 static void
494 svc_vc_backchannel_destroy(SVCXPRT *xprt)
495 {
496 	struct cf_conn *cd = (struct cf_conn *)xprt->xp_p1;
497 	struct mbuf *m, *m2;
498 
499 	svc_xprt_free(xprt);
500 	m = cd->mreq;
501 	while (m != NULL) {
502 		m2 = m;
503 		m = m->m_nextpkt;
504 		m_freem(m2);
505 	}
506 	mem_free(cd, sizeof(*cd));
507 }
508 
509 /*ARGSUSED*/
510 static bool_t
511 svc_vc_control(SVCXPRT *xprt, const u_int rq, void *in)
512 {
513 	return (FALSE);
514 }
515 
516 static bool_t
517 svc_vc_rendezvous_control(SVCXPRT *xprt, const u_int rq, void *in)
518 {
519 
520 	return (FALSE);
521 }
522 
523 static bool_t
524 svc_vc_backchannel_control(SVCXPRT *xprt, const u_int rq, void *in)
525 {
526 
527 	return (FALSE);
528 }
529 
530 static enum xprt_stat
531 svc_vc_stat(SVCXPRT *xprt)
532 {
533 	struct cf_conn *cd;
534 
535 	cd = (struct cf_conn *)(xprt->xp_p1);
536 
537 	if (cd->strm_stat == XPRT_DIED)
538 		return (XPRT_DIED);
539 
540 	if (cd->mreq != NULL && cd->resid == 0 && cd->eor)
541 		return (XPRT_MOREREQS);
542 
543 	if (soreadable(xprt->xp_socket))
544 		return (XPRT_MOREREQS);
545 
546 	return (XPRT_IDLE);
547 }
548 
549 static bool_t
550 svc_vc_ack(SVCXPRT *xprt, uint32_t *ack)
551 {
552 
553 	*ack = atomic_load_acq_32(&xprt->xp_snt_cnt);
554 	*ack -= sbused(&xprt->xp_socket->so_snd);
555 	return (TRUE);
556 }
557 
558 static enum xprt_stat
559 svc_vc_backchannel_stat(SVCXPRT *xprt)
560 {
561 	struct cf_conn *cd;
562 
563 	cd = (struct cf_conn *)(xprt->xp_p1);
564 
565 	if (cd->mreq != NULL)
566 		return (XPRT_MOREREQS);
567 
568 	return (XPRT_IDLE);
569 }
570 
571 /*
572  * If we have an mbuf chain in cd->mpending, try to parse a record from it,
573  * leaving the result in cd->mreq. If we don't have a complete record, leave
574  * the partial result in cd->mreq and try to read more from the socket.
575  */
576 static int
577 svc_vc_process_pending(SVCXPRT *xprt)
578 {
579 	struct cf_conn *cd = (struct cf_conn *) xprt->xp_p1;
580 	struct socket *so = xprt->xp_socket;
581 	struct mbuf *m;
582 
583 	/*
584 	 * If cd->resid is non-zero, we have part of the
585 	 * record already, otherwise we are expecting a record
586 	 * marker.
587 	 */
588 	if (!cd->resid && cd->mpending) {
589 		/*
590 		 * See if there is enough data buffered to
591 		 * make up a record marker. Make sure we can
592 		 * handle the case where the record marker is
593 		 * split across more than one mbuf.
594 		 */
595 		size_t n = 0;
596 		uint32_t header;
597 
598 		m = cd->mpending;
599 		while (n < sizeof(uint32_t) && m) {
600 			n += m->m_len;
601 			m = m->m_next;
602 		}
603 		if (n < sizeof(uint32_t)) {
604 			so->so_rcv.sb_lowat = sizeof(uint32_t) - n;
605 			return (FALSE);
606 		}
607 		m_copydata(cd->mpending, 0, sizeof(header),
608 		    (char *)&header);
609 		header = ntohl(header);
610 		cd->eor = (header & 0x80000000) != 0;
611 		cd->resid = header & 0x7fffffff;
612 		m_adj(cd->mpending, sizeof(uint32_t));
613 	}
614 
615 	/*
616 	 * Start pulling off mbufs from cd->mpending
617 	 * until we either have a complete record or
618 	 * we run out of data. We use m_split to pull
619 	 * data - it will pull as much as possible and
620 	 * split the last mbuf if necessary.
621 	 */
622 	while (cd->mpending && cd->resid) {
623 		m = cd->mpending;
624 		if (cd->mpending->m_next
625 		    || cd->mpending->m_len > cd->resid)
626 			cd->mpending = m_split(cd->mpending,
627 			    cd->resid, M_WAITOK);
628 		else
629 			cd->mpending = NULL;
630 		if (cd->mreq)
631 			m_last(cd->mreq)->m_next = m;
632 		else
633 			cd->mreq = m;
634 		while (m) {
635 			cd->resid -= m->m_len;
636 			m = m->m_next;
637 		}
638 	}
639 
640 	/*
641 	 * Block receive upcalls if we have more data pending,
642 	 * otherwise report our need.
643 	 */
644 	if (cd->mpending)
645 		so->so_rcv.sb_lowat = INT_MAX;
646 	else
647 		so->so_rcv.sb_lowat =
648 		    imax(1, imin(cd->resid, so->so_rcv.sb_hiwat / 2));
649 	return (TRUE);
650 }
651 
652 static bool_t
653 svc_vc_recv(SVCXPRT *xprt, struct rpc_msg *msg,
654     struct sockaddr **addrp, struct mbuf **mp)
655 {
656 	struct cf_conn *cd = (struct cf_conn *) xprt->xp_p1;
657 	struct uio uio;
658 	struct mbuf *m;
659 	struct socket* so = xprt->xp_socket;
660 	XDR xdrs;
661 	int error, rcvflag;
662 	uint32_t xid_plus_direction[2];
663 
664 	/*
665 	 * Serialise access to the socket and our own record parsing
666 	 * state.
667 	 */
668 	sx_xlock(&xprt->xp_lock);
669 
670 	for (;;) {
671 		/* If we have no request ready, check pending queue. */
672 		while (cd->mpending &&
673 		    (cd->mreq == NULL || cd->resid != 0 || !cd->eor)) {
674 			if (!svc_vc_process_pending(xprt))
675 				break;
676 		}
677 
678 		/* Process and return complete request in cd->mreq. */
679 		if (cd->mreq != NULL && cd->resid == 0 && cd->eor) {
680 
681 			/*
682 			 * Now, check for a backchannel reply.
683 			 * The XID is in the first uint32_t of the reply
684 			 * and the message direction is the second one.
685 			 */
686 			if ((cd->mreq->m_len >= sizeof(xid_plus_direction) ||
687 			    m_length(cd->mreq, NULL) >=
688 			    sizeof(xid_plus_direction)) &&
689 			    xprt->xp_p2 != NULL) {
690 				m_copydata(cd->mreq, 0,
691 				    sizeof(xid_plus_direction),
692 				    (char *)xid_plus_direction);
693 				xid_plus_direction[0] =
694 				    ntohl(xid_plus_direction[0]);
695 				xid_plus_direction[1] =
696 				    ntohl(xid_plus_direction[1]);
697 				/* Check message direction. */
698 				if (xid_plus_direction[1] == REPLY) {
699 					clnt_bck_svccall(xprt->xp_p2,
700 					    cd->mreq,
701 					    xid_plus_direction[0]);
702 					cd->mreq = NULL;
703 					continue;
704 				}
705 			}
706 
707 			xdrmbuf_create(&xdrs, cd->mreq, XDR_DECODE);
708 			cd->mreq = NULL;
709 
710 			/* Check for next request in a pending queue. */
711 			svc_vc_process_pending(xprt);
712 			if (cd->mreq == NULL || cd->resid != 0) {
713 				SOCKBUF_LOCK(&so->so_rcv);
714 				if (!soreadable(so))
715 					xprt_inactive_self(xprt);
716 				SOCKBUF_UNLOCK(&so->so_rcv);
717 			}
718 
719 			sx_xunlock(&xprt->xp_lock);
720 
721 			if (! xdr_callmsg(&xdrs, msg)) {
722 				XDR_DESTROY(&xdrs);
723 				return (FALSE);
724 			}
725 
726 			*addrp = NULL;
727 			*mp = xdrmbuf_getall(&xdrs);
728 			XDR_DESTROY(&xdrs);
729 
730 			return (TRUE);
731 		}
732 
733 		/*
734 		 * The socket upcall calls xprt_active() which will eventually
735 		 * cause the server to call us here. We attempt to
736 		 * read as much as possible from the socket and put
737 		 * the result in cd->mpending. If the read fails,
738 		 * we have drained both cd->mpending and the socket so
739 		 * we can call xprt_inactive().
740 		 */
741 		uio.uio_resid = 1000000000;
742 		uio.uio_td = curthread;
743 		m = NULL;
744 		rcvflag = MSG_DONTWAIT;
745 		error = soreceive(so, NULL, &uio, &m, NULL, &rcvflag);
746 
747 		if (error == EWOULDBLOCK) {
748 			/*
749 			 * We must re-test for readability after
750 			 * taking the lock to protect us in the case
751 			 * where a new packet arrives on the socket
752 			 * after our call to soreceive fails with
753 			 * EWOULDBLOCK.
754 			 */
755 			SOCKBUF_LOCK(&so->so_rcv);
756 			if (!soreadable(so))
757 				xprt_inactive_self(xprt);
758 			SOCKBUF_UNLOCK(&so->so_rcv);
759 			sx_xunlock(&xprt->xp_lock);
760 			return (FALSE);
761 		}
762 
763 		if (error) {
764 			SOCKBUF_LOCK(&so->so_rcv);
765 			if (xprt->xp_upcallset) {
766 				xprt->xp_upcallset = 0;
767 				soupcall_clear(so, SO_RCV);
768 			}
769 			SOCKBUF_UNLOCK(&so->so_rcv);
770 			xprt_inactive_self(xprt);
771 			cd->strm_stat = XPRT_DIED;
772 			sx_xunlock(&xprt->xp_lock);
773 			return (FALSE);
774 		}
775 
776 		if (!m) {
777 			/*
778 			 * EOF - the other end has closed the socket.
779 			 */
780 			xprt_inactive_self(xprt);
781 			cd->strm_stat = XPRT_DIED;
782 			sx_xunlock(&xprt->xp_lock);
783 			return (FALSE);
784 		}
785 
786 		if (cd->mpending)
787 			m_last(cd->mpending)->m_next = m;
788 		else
789 			cd->mpending = m;
790 	}
791 }
792 
793 static bool_t
794 svc_vc_backchannel_recv(SVCXPRT *xprt, struct rpc_msg *msg,
795     struct sockaddr **addrp, struct mbuf **mp)
796 {
797 	struct cf_conn *cd = (struct cf_conn *) xprt->xp_p1;
798 	struct ct_data *ct;
799 	struct mbuf *m;
800 	XDR xdrs;
801 
802 	sx_xlock(&xprt->xp_lock);
803 	ct = (struct ct_data *)xprt->xp_p2;
804 	if (ct == NULL) {
805 		sx_xunlock(&xprt->xp_lock);
806 		return (FALSE);
807 	}
808 	mtx_lock(&ct->ct_lock);
809 	m = cd->mreq;
810 	if (m == NULL) {
811 		xprt_inactive_self(xprt);
812 		mtx_unlock(&ct->ct_lock);
813 		sx_xunlock(&xprt->xp_lock);
814 		return (FALSE);
815 	}
816 	cd->mreq = m->m_nextpkt;
817 	mtx_unlock(&ct->ct_lock);
818 	sx_xunlock(&xprt->xp_lock);
819 
820 	xdrmbuf_create(&xdrs, m, XDR_DECODE);
821 	if (! xdr_callmsg(&xdrs, msg)) {
822 		XDR_DESTROY(&xdrs);
823 		return (FALSE);
824 	}
825 	*addrp = NULL;
826 	*mp = xdrmbuf_getall(&xdrs);
827 	XDR_DESTROY(&xdrs);
828 	return (TRUE);
829 }
830 
831 static bool_t
832 svc_vc_reply(SVCXPRT *xprt, struct rpc_msg *msg,
833     struct sockaddr *addr, struct mbuf *m, uint32_t *seq)
834 {
835 	XDR xdrs;
836 	struct mbuf *mrep;
837 	bool_t stat = TRUE;
838 	int error, len;
839 
840 	/*
841 	 * Leave space for record mark.
842 	 */
843 	mrep = m_gethdr(M_WAITOK, MT_DATA);
844 	mrep->m_data += sizeof(uint32_t);
845 
846 	xdrmbuf_create(&xdrs, mrep, XDR_ENCODE);
847 
848 	if (msg->rm_reply.rp_stat == MSG_ACCEPTED &&
849 	    msg->rm_reply.rp_acpt.ar_stat == SUCCESS) {
850 		if (!xdr_replymsg(&xdrs, msg))
851 			stat = FALSE;
852 		else
853 			xdrmbuf_append(&xdrs, m);
854 	} else {
855 		stat = xdr_replymsg(&xdrs, msg);
856 	}
857 
858 	if (stat) {
859 		m_fixhdr(mrep);
860 
861 		/*
862 		 * Prepend a record marker containing the reply length.
863 		 */
864 		M_PREPEND(mrep, sizeof(uint32_t), M_WAITOK);
865 		len = mrep->m_pkthdr.len;
866 		*mtod(mrep, uint32_t *) =
867 			htonl(0x80000000 | (len - sizeof(uint32_t)));
868 		atomic_add_32(&xprt->xp_snd_cnt, len);
869 		error = sosend(xprt->xp_socket, NULL, NULL, mrep, NULL,
870 		    0, curthread);
871 		if (!error) {
872 			atomic_add_rel_32(&xprt->xp_snt_cnt, len);
873 			if (seq)
874 				*seq = xprt->xp_snd_cnt;
875 			stat = TRUE;
876 		} else
877 			atomic_subtract_32(&xprt->xp_snd_cnt, len);
878 	} else {
879 		m_freem(mrep);
880 	}
881 
882 	XDR_DESTROY(&xdrs);
883 
884 	return (stat);
885 }
886 
887 static bool_t
888 svc_vc_backchannel_reply(SVCXPRT *xprt, struct rpc_msg *msg,
889     struct sockaddr *addr, struct mbuf *m, uint32_t *seq)
890 {
891 	struct ct_data *ct;
892 	XDR xdrs;
893 	struct mbuf *mrep;
894 	bool_t stat = TRUE;
895 	int error;
896 
897 	/*
898 	 * Leave space for record mark.
899 	 */
900 	mrep = m_gethdr(M_WAITOK, MT_DATA);
901 	mrep->m_data += sizeof(uint32_t);
902 
903 	xdrmbuf_create(&xdrs, mrep, XDR_ENCODE);
904 
905 	if (msg->rm_reply.rp_stat == MSG_ACCEPTED &&
906 	    msg->rm_reply.rp_acpt.ar_stat == SUCCESS) {
907 		if (!xdr_replymsg(&xdrs, msg))
908 			stat = FALSE;
909 		else
910 			xdrmbuf_append(&xdrs, m);
911 	} else {
912 		stat = xdr_replymsg(&xdrs, msg);
913 	}
914 
915 	if (stat) {
916 		m_fixhdr(mrep);
917 
918 		/*
919 		 * Prepend a record marker containing the reply length.
920 		 */
921 		M_PREPEND(mrep, sizeof(uint32_t), M_WAITOK);
922 		*mtod(mrep, uint32_t *) =
923 			htonl(0x80000000 | (mrep->m_pkthdr.len
924 				- sizeof(uint32_t)));
925 		sx_xlock(&xprt->xp_lock);
926 		ct = (struct ct_data *)xprt->xp_p2;
927 		if (ct != NULL)
928 			error = sosend(ct->ct_socket, NULL, NULL, mrep, NULL,
929 			    0, curthread);
930 		else
931 			error = EPIPE;
932 		sx_xunlock(&xprt->xp_lock);
933 		if (!error) {
934 			stat = TRUE;
935 		}
936 	} else {
937 		m_freem(mrep);
938 	}
939 
940 	XDR_DESTROY(&xdrs);
941 
942 	return (stat);
943 }
944 
945 static bool_t
946 svc_vc_null()
947 {
948 
949 	return (FALSE);
950 }
951 
952 static int
953 svc_vc_soupcall(struct socket *so, void *arg, int waitflag)
954 {
955 	SVCXPRT *xprt = (SVCXPRT *) arg;
956 
957 	if (soreadable(xprt->xp_socket))
958 		xprt_active(xprt);
959 	return (SU_OK);
960 }
961 
962 static int
963 svc_vc_rendezvous_soupcall(struct socket *head, void *arg, int waitflag)
964 {
965 	SVCXPRT *xprt = (SVCXPRT *) arg;
966 
967 	if (!TAILQ_EMPTY(&head->sol_comp))
968 		xprt_active(xprt);
969 	return (SU_OK);
970 }
971 
972 #if 0
973 /*
974  * Get the effective UID of the sending process. Used by rpcbind, keyserv
975  * and rpc.yppasswdd on AF_LOCAL.
976  */
977 int
978 __rpc_get_local_uid(SVCXPRT *transp, uid_t *uid) {
979 	int sock, ret;
980 	gid_t egid;
981 	uid_t euid;
982 	struct sockaddr *sa;
983 
984 	sock = transp->xp_fd;
985 	sa = (struct sockaddr *)transp->xp_rtaddr;
986 	if (sa->sa_family == AF_LOCAL) {
987 		ret = getpeereid(sock, &euid, &egid);
988 		if (ret == 0)
989 			*uid = euid;
990 		return (ret);
991 	} else
992 		return (-1);
993 }
994 #endif
995